Mercurial > hg > sv-dependency-builds

comparison src/fftw-3.3.8/rdft/scalar/r2cf/r2cf_14.c @ 167:bd3cc4d1df30

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add FFTW 3.3.8 source, and a Linux build
author Chris Cannam <cannam@all-day-breakfast.com>
date Tue, 19 Nov 2019 14:52:55 +0000
parents
children
comparison
equal deleted inserted replaced
166:cbd6d7e562c7 167:bd3cc4d1df30
1 /*
2 * Copyright (c) 2003, 2007-14 Matteo Frigo
3 * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
4 *
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
9 *
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
14 *
15 * You should have received a copy of the GNU General Public License
16 * along with this program; if not, write to the Free Software
17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
18 *
19 */
20
21 /* This file was automatically generated --- DO NOT EDIT */
22 /* Generated on Thu May 24 08:06:26 EDT 2018 */
23
24 #include "rdft/codelet-rdft.h"
25
26 #if defined(ARCH_PREFERS_FMA) || defined(ISA_EXTENSION_PREFERS_FMA)
27
28 /* Generated by: ../../../genfft/gen_r2cf.native -fma -compact -variables 4 -pipeline-latency 4 -n 14 -name r2cf_14 -include rdft/scalar/r2cf.h */
29
30 /*
31 * This function contains 62 FP additions, 36 FP multiplications,
32 * (or, 32 additions, 6 multiplications, 30 fused multiply/add),
33 * 33 stack variables, 6 constants, and 28 memory accesses
34 */
35 #include "rdft/scalar/r2cf.h"
36
37 static void r2cf_14(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
38 {
39 DK(KP900968867, +0.900968867902419126236102319507445051165919162);
40 DK(KP692021471, +0.692021471630095869627814897002069140197260599);
41 DK(KP356895867, +0.356895867892209443894399510021300583399127187);
42 DK(KP801937735, +0.801937735804838252472204639014890102331838324);
43 DK(KP974927912, +0.974927912181823607018131682993931217232785801);
44 DK(KP554958132, +0.554958132087371191422194871006410481067288862);
45 {
46 INT i;
47 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(56, rs), MAKE_VOLATILE_STRIDE(56, csr), MAKE_VOLATILE_STRIDE(56, csi)) {
48 E T3, TN, To, TQ, Tx, TG, Ta, TO, Tw, TD, Th, TP, Tv, TJ, T1;
49 E T2, TA, TK;
50 T1 = R0[0];
51 T2 = R1[WS(rs, 3)];
52 T3 = T1 - T2;
53 TN = T1 + T2;
54 {
55 E Tk, TE, Tn, TF;
56 {
57 E Ti, Tj, Tl, Tm;
58 Ti = R0[WS(rs, 3)];
59 Tj = R1[WS(rs, 6)];
60 Tk = Ti - Tj;
61 TE = Ti + Tj;
62 Tl = R0[WS(rs, 4)];
63 Tm = R1[0];
64 Tn = Tl - Tm;
65 TF = Tl + Tm;
66 }
67 To = Tk + Tn;
68 TQ = TE + TF;
69 Tx = Tn - Tk;
70 TG = TE - TF;
71 }
72 {
73 E T6, TC, T9, TB;
74 {
75 E T4, T5, T7, T8;
76 T4 = R0[WS(rs, 1)];
77 T5 = R1[WS(rs, 4)];
78 T6 = T4 - T5;
79 TC = T4 + T5;
80 T7 = R0[WS(rs, 6)];
81 T8 = R1[WS(rs, 2)];
82 T9 = T7 - T8;
83 TB = T7 + T8;
84 }
85 Ta = T6 + T9;
86 TO = TC + TB;
87 Tw = T6 - T9;
88 TD = TB - TC;
89 }
90 {
91 E Td, TH, Tg, TI;
92 {
93 E Tb, Tc, Te, Tf;
94 Tb = R0[WS(rs, 2)];
95 Tc = R1[WS(rs, 5)];
96 Td = Tb - Tc;
97 TH = Tb + Tc;
98 Te = R0[WS(rs, 5)];
99 Tf = R1[WS(rs, 1)];
100 Tg = Te - Tf;
101 TI = Te + Tf;
102 }
103 Th = Td + Tg;
104 TP = TH + TI;
105 Tv = Tg - Td;
106 TJ = TH - TI;
107 }
108 Cr[WS(csr, 7)] = T3 + Ta + Th + To;
109 Cr[0] = TN + TO + TP + TQ;
110 TA = FMA(KP554958132, Tw, Tv);
111 Ci[WS(csi, 3)] = KP974927912 * (FNMS(KP801937735, TA, Tx));
112 {
113 E TL, TM, Ty, Tz;
114 TL = FNMS(KP554958132, TG, TD);
115 Ci[WS(csi, 6)] = KP974927912 * (FNMS(KP801937735, TL, TJ));
116 TM = FMA(KP554958132, TD, TJ);
117 Ci[WS(csi, 4)] = KP974927912 * (FNMS(KP801937735, TM, TG));
118 Ty = FNMS(KP554958132, Tx, Tw);
119 Ci[WS(csi, 1)] = KP974927912 * (FNMS(KP801937735, Ty, Tv));
120 Tz = FMA(KP554958132, Tv, Tx);
121 Ci[WS(csi, 5)] = KP974927912 * (FMA(KP801937735, Tz, Tw));
122 }
123 TK = FMA(KP554958132, TJ, TG);
124 Ci[WS(csi, 2)] = KP974927912 * (FMA(KP801937735, TK, TD));
125 {
126 E TU, TT, Tq, Tp;
127 TT = FNMS(KP356895867, TO, TQ);
128 TU = FNMS(KP692021471, TT, TP);
129 Cr[WS(csr, 2)] = FNMS(KP900968867, TU, TN);
130 Tp = FNMS(KP356895867, To, Th);
131 Tq = FNMS(KP692021471, Tp, Ta);
132 Cr[WS(csr, 3)] = FNMS(KP900968867, Tq, T3);
133 }
134 {
135 E Tu, Tt, Ts, Tr;
136 Tt = FNMS(KP356895867, Th, Ta);
137 Tu = FNMS(KP692021471, Tt, To);
138 Cr[WS(csr, 1)] = FNMS(KP900968867, Tu, T3);
139 Tr = FNMS(KP356895867, Ta, To);
140 Ts = FNMS(KP692021471, Tr, Th);
141 Cr[WS(csr, 5)] = FNMS(KP900968867, Ts, T3);
142 }
143 {
144 E TW, TV, TS, TR;
145 TV = FNMS(KP356895867, TP, TO);
146 TW = FNMS(KP692021471, TV, TQ);
147 Cr[WS(csr, 6)] = FNMS(KP900968867, TW, TN);
148 TR = FNMS(KP356895867, TQ, TP);
149 TS = FNMS(KP692021471, TR, TO);
150 Cr[WS(csr, 4)] = FNMS(KP900968867, TS, TN);
151 }
152 }
153 }
154 }
155
156 static const kr2c_desc desc = { 14, "r2cf_14", {32, 6, 30, 0}, &GENUS };
157
158 void X(codelet_r2cf_14) (planner *p) {
159 X(kr2c_register) (p, r2cf_14, &desc);
160 }
161
162 #else
163
164 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 14 -name r2cf_14 -include rdft/scalar/r2cf.h */
165
166 /*
167 * This function contains 62 FP additions, 36 FP multiplications,
168 * (or, 38 additions, 12 multiplications, 24 fused multiply/add),
169 * 29 stack variables, 6 constants, and 28 memory accesses
170 */
171 #include "rdft/scalar/r2cf.h"
172
173 static void r2cf_14(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
174 {
175 DK(KP900968867, +0.900968867902419126236102319507445051165919162);
176 DK(KP222520933, +0.222520933956314404288902564496794759466355569);
177 DK(KP623489801, +0.623489801858733530525004884004239810632274731);
178 DK(KP433883739, +0.433883739117558120475768332848358754609990728);
179 DK(KP974927912, +0.974927912181823607018131682993931217232785801);
180 DK(KP781831482, +0.781831482468029808708444526674057750232334519);
181 {
182 INT i;
183 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(56, rs), MAKE_VOLATILE_STRIDE(56, csr), MAKE_VOLATILE_STRIDE(56, csi)) {
184 E T3, TB, T6, Tv, Tn, Ts, Tk, Tt, Td, Ty, T9, Tw, Tg, Tz, T1;
185 E T2;
186 T1 = R0[0];
187 T2 = R1[WS(rs, 3)];
188 T3 = T1 - T2;
189 TB = T1 + T2;
190 {
191 E T4, T5, Tl, Tm;
192 T4 = R0[WS(rs, 2)];
193 T5 = R1[WS(rs, 5)];
194 T6 = T4 - T5;
195 Tv = T4 + T5;
196 Tl = R0[WS(rs, 6)];
197 Tm = R1[WS(rs, 2)];
198 Tn = Tl - Tm;
199 Ts = Tl + Tm;
200 }
201 {
202 E Ti, Tj, Tb, Tc;
203 Ti = R0[WS(rs, 1)];
204 Tj = R1[WS(rs, 4)];
205 Tk = Ti - Tj;
206 Tt = Ti + Tj;
207 Tb = R0[WS(rs, 3)];
208 Tc = R1[WS(rs, 6)];
209 Td = Tb - Tc;
210 Ty = Tb + Tc;
211 }
212 {
213 E T7, T8, Te, Tf;
214 T7 = R0[WS(rs, 5)];
215 T8 = R1[WS(rs, 1)];
216 T9 = T7 - T8;
217 Tw = T7 + T8;
218 Te = R0[WS(rs, 4)];
219 Tf = R1[0];
220 Tg = Te - Tf;
221 Tz = Te + Tf;
222 }
223 {
224 E Tp, Tr, Tq, Ta, To, Th;
225 Tp = Tn - Tk;
226 Tr = Tg - Td;
227 Tq = T9 - T6;
228 Ci[WS(csi, 1)] = FMA(KP781831482, Tp, KP974927912 * Tq) + (KP433883739 * Tr);
229 Ci[WS(csi, 5)] = FMA(KP433883739, Tq, KP781831482 * Tr) - (KP974927912 * Tp);
230 Ci[WS(csi, 3)] = FMA(KP433883739, Tp, KP974927912 * Tr) - (KP781831482 * Tq);
231 Ta = T6 + T9;
232 To = Tk + Tn;
233 Th = Td + Tg;
234 Cr[WS(csr, 3)] = FMA(KP623489801, Ta, T3) + FNMA(KP222520933, Th, KP900968867 * To);
235 Cr[WS(csr, 7)] = T3 + To + Ta + Th;
236 Cr[WS(csr, 1)] = FMA(KP623489801, To, T3) + FNMA(KP900968867, Th, KP222520933 * Ta);
237 Cr[WS(csr, 5)] = FMA(KP623489801, Th, T3) + FNMA(KP900968867, Ta, KP222520933 * To);
238 }
239 {
240 E Tu, TA, Tx, TC, TE, TD;
241 Tu = Ts - Tt;
242 TA = Ty - Tz;
243 Tx = Tv - Tw;
244 Ci[WS(csi, 2)] = FMA(KP974927912, Tu, KP433883739 * Tx) + (KP781831482 * TA);
245 Ci[WS(csi, 6)] = FMA(KP974927912, Tx, KP433883739 * TA) - (KP781831482 * Tu);
246 Ci[WS(csi, 4)] = FNMS(KP781831482, Tx, KP974927912 * TA) - (KP433883739 * Tu);
247 TC = Tt + Ts;
248 TE = Tv + Tw;
249 TD = Ty + Tz;
250 Cr[WS(csr, 6)] = FMA(KP623489801, TC, TB) + FNMA(KP900968867, TD, KP222520933 * TE);
251 Cr[WS(csr, 2)] = FMA(KP623489801, TD, TB) + FNMA(KP900968867, TE, KP222520933 * TC);
252 Cr[WS(csr, 4)] = FMA(KP623489801, TE, TB) + FNMA(KP222520933, TD, KP900968867 * TC);
253 Cr[0] = TB + TC + TE + TD;
254 }
255 }
256 }
257 }
258
259 static const kr2c_desc desc = { 14, "r2cf_14", {38, 12, 24, 0}, &GENUS };
260
261 void X(codelet_r2cf_14) (planner *p) {
262 X(kr2c_register) (p, r2cf_14, &desc);
263 }
264
265 #endif